Joe L.

Yes. Quite normal. A 2TB drive usually takes between 33 and 36 hours depending on what else is going on on the server. If you are doing a lot of other disk IO, it could take a bit longer. Yes, also normal. The inner cylinders on the disk do not hold as much data as the outer ones because their tracks are not as long, since the rotational speed is constant, the potential "read rate" of inner cylinders is always slower. Let's assume an average "read" speed of 75 MB/s. That means you'll read 1000 MB (1 GB) in 13.333 seconds. We need to read 2000 GB (2TB) so, we would get at best 13.333 * 2000 seconds = 26,666.666 seconds 26,666 seconds = 444.44 minutes 444.44 minutes = 7.4 hours. So, 7.4 hours roughly just to read the entire disk, and you cannot speed that up no matter what you do, and that does not count the time verifying the zeros were written properly in the post read phase. Joe L.

Nothing looks weird to me. What looks weird to you? I thought that > 241 Unknown_Attribute 0x0032 200 200 000 Old_age Always - 625144849 > 242 Unknown_Attribute 0x0032 200 200 000 Old_age Always - 1903926051 was weird, but if there is nothing to worry about I'm cool. Thanks. It indicates the manufacturer has not told anybody about what those two parameters represent, but that their "normalized" value of 200 is nowhere near their failure threshold of "000" Joe L.

About all we care about is re-allocated sectors and sectors pending re-allocation. Of course, if any given parameter's "normalized" value is getting near its affiliated failure threshold that would be a cause for concern.

Nothing looks weird to me. What looks weird to you?

Looks good to me too.

The two SMART reports were logged to your syslog. You can scroll up on the system console... (Shift-PageUp and Shift-PageDown) Joe L.

looks like communications with the drive is failing. Might try a different SATA cable, or re-seat the existing cable to the drive. You have lots of BadCRC errors. Aug 4 12:57:08 Tower kernel: ata1.00: exception Emask 0x50 SAct 0x3 SErr 0x280900 action 0x6 frozen Aug 4 12:57:08 Tower kernel: ata1.00: irq_stat 0x08000000, interface fatal error Aug 4 12:57:08 Tower kernel: ata1: SError: { UnrecovData HostInt 10B8B BadCRC } Aug 4 12:57:08 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:08 Tower kernel: ata1.00: cmd 60/00:00:b8:24:1d/01:00:09:00:00/40 tag 0 ncq 131072 in Aug 4 12:57:08 Tower kernel: res 40/00:0c:b8:23:1d/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:08 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:08 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:08 Tower kernel: ata1.00: cmd 60/00:08:b8:23:1d/01:00:09:00:00/40 tag 1 ncq 131072 in Aug 4 12:57:08 Tower kernel: res 40/00:0c:b8:23:1d/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:08 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:08 Tower kernel: ata1: hard resetting link Aug 4 12:57:08 Tower kernel: ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Aug 4 12:57:08 Tower kernel: ata1.00: configured for UDMA/133 Aug 4 12:57:08 Tower kernel: ata1: EH complete Aug 4 12:57:09 Tower kernel: ata1.00: exception Emask 0x50 SAct 0x3 SErr 0x280900 action 0x6 frozen Aug 4 12:57:09 Tower kernel: ata1.00: irq_stat 0x08000000, interface fatal error Aug 4 12:57:09 Tower kernel: ata1: SError: { UnrecovData HostInt 10B8B BadCRC } Aug 4 12:57:09 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:09 Tower kernel: ata1.00: cmd 60/00:00:b8:4d:20/01:00:09:00:00/40 tag 0 ncq 131072 in Aug 4 12:57:09 Tower kernel: res 40/00:0c:b8:4c:20/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:09 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:09 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:09 Tower kernel: ata1.00: cmd 60/00:08:b8:4c:20/01:00:09:00:00/40 tag 1 ncq 131072 in Aug 4 12:57:09 Tower kernel: res 40/00:0c:b8:4c:20/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:09 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:09 Tower kernel: ata1: hard resetting link Aug 4 12:57:10 Tower kernel: ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Aug 4 12:57:10 Tower kernel: ata1.00: configured for UDMA/133 Aug 4 12:57:10 Tower kernel: ata1: EH complete Aug 4 12:57:10 Tower kernel: ata1.00: exception Emask 0x50 SAct 0x3 SErr 0x280900 action 0x6 frozen Aug 4 12:57:10 Tower kernel: ata1.00: irq_stat 0x08000000, interface fatal error Aug 4 12:57:10 Tower kernel: ata1: SError: { UnrecovData HostInt 10B8B BadCRC } Aug 4 12:57:10 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:10 Tower kernel: ata1.00: cmd 60/00:00:b8:a2:20/01:00:09:00:00/40 tag 0 ncq 131072 in Aug 4 12:57:10 Tower kernel: res 40/00:04:b8:a2:20/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:10 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:10 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:10 Tower kernel: ata1.00: cmd 60/00:08:b8:a3:20/01:00:09:00:00/40 tag 1 ncq 131072 in Aug 4 12:57:10 Tower kernel: res 40/00:04:b8:a2:20/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:10 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:10 Tower kernel: ata1: hard resetting link Aug 4 12:57:11 Tower kernel: ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Aug 4 12:57:11 Tower kernel: ata1.00: configured for UDMA/133 Aug 4 12:57:11 Tower kernel: ata1: EH complete Aug 4 12:57:11 Tower kernel: ata1: limiting SATA link speed to 1.5 Gbps Aug 4 12:57:11 Tower kernel: ata1.00: exception Emask 0x50 SAct 0x3 SErr 0x280900 action 0x6 frozen Aug 4 12:57:11 Tower kernel: ata1.00: irq_stat 0x08000000, interface fatal error Aug 4 12:57:11 Tower kernel: ata1: SError: { UnrecovData HostInt 10B8B BadCRC } Aug 4 12:57:11 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:11 Tower kernel: ata1.00: cmd 60/00:00:b8:db:20/01:00:09:00:00/40 tag 0 ncq 131072 in Aug 4 12:57:11 Tower kernel: res 40/00:0c:b8:dc:20/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:11 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:11 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:11 Tower kernel: ata1.00: cmd 60/00:08:b8:dc:20/01:00:09:00:00/40 tag 1 ncq 131072 in Aug 4 12:57:11 Tower kernel: res 40/00:0c:b8:dc:20/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:11 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:11 Tower kernel: ata1: hard resetting link Aug 4 12:57:11 Tower kernel: ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Aug 4 12:57:11 Tower kernel: ata1.00: configured for UDMA/133 Aug 4 12:57:11 Tower kernel: ata1: EH complete Aug 4 12:57:11 Tower kernel: ata1.00: exception Emask 0x50 SAct 0x3 SErr 0x280900 action 0x6 frozen Aug 4 12:57:11 Tower kernel: ata1.00: irq_stat 0x08000000, interface fatal error Aug 4 12:57:11 Tower kernel: ata1: SError: { UnrecovData HostInt 10B8B BadCRC } Aug 4 12:57:11 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:11 Tower kernel: ata1.00: cmd 60/00:00:b8:dc:20/01:00:09:00:00/40 tag 0 ncq 131072 in Aug 4 12:57:11 Tower kernel: res 40/00:0c:b8:db:20/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:11 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:11 Tower kernel: ata1.00: failed command: READ FPDMA QUEUED Aug 4 12:57:11 Tower kernel: ata1.00: cmd 60/00:08:b8:db:20/01:00:09:00:00/40 tag 1 ncq 131072 in Aug 4 12:57:11 Tower kernel: res 40/00:0c:b8:db:20/00:00:09:00:00/40 Emask 0x50 (ATA bus error) Aug 4 12:57:11 Tower kernel: ata1.00: status: { DRDY } Aug 4 12:57:11 Tower kernel: ata1: hard resetting link Aug 4 12:57:12 Tower kernel: ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Aug 4 12:57:12 Tower kernel: ata1.00: failed to IDENTIFY (I/O error, err_mask=0x100) Aug 4 12:57:12 Tower kernel: ata1.00: revalidation failed (errno=-5) Aug 4 12:57:17 Tower kernel: ata1: hard resetting link Aug 4 12:57:17 Tower kernel: ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Aug 4 12:57:17 Tower kernel: ata1.00: failed to IDENTIFY (I/O error, err_mask=0x100) Aug 4 12:57:17 Tower kernel: ata1.00: revalidation failed (errno=-5) Aug 4 12:57:22 Tower kernel: ata1: hard resetting link Aug 4 12:57:23 Tower kernel: ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Aug 4 12:57:23 Tower kernel: ata1.00: failed to IDENTIFY (I/O error, err_mask=0x100) Aug 4 12:57:23 Tower kernel: ata1.00: revalidation failed (errno=-5) Aug 4 12:57:23 Tower kernel: ata1.00: disabled Aug 4 12:57:23 Tower kernel: ata1: exception Emask 0x52 SAct 0x0 SErr 0x280d00 action 0x6 frozen t4 Aug 4 12:57:23 Tower kernel: ata1: irq_stat 0x08000000, interface fatal error Aug 4 12:57:23 Tower kernel: ata1: SError: { UnrecovData Proto HostInt 10B8B BadCRC } Aug 4 12:57:23 Tower kernel: ata1: hard resetting link Aug 4 12:57:23 Tower kernel: ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Aug 4 12:57:23 Tower kernel: ata1: EH complete Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Unhandled error code Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] CDB: cdb[0]=0x28: 28 00 09 20 db b8 00 01 00 00 Aug 4 12:57:23 Tower kernel: end_request: I/O error, dev sda, sector 153148344 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143543 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143544 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143545 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143546 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143547 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143548 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143549 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143550 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143551 Aug 4 12:57:23 Tower kernel: Buffer I/O error on device sda, logical block 19143552 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Unhandled error code Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] CDB: cdb[0]=0x28: 28 00 09 20 dc b8 00 01 00 00 Aug 4 12:57:23 Tower kernel: end_request: I/O error, dev sda, sector 153148600 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Unhandled error code Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] CDB: cdb[0]=0x28: 28 00 09 20 db b8 00 00 08 00 Aug 4 12:57:23 Tower kernel: end_request: I/O error, dev sda, sector 153148344 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Unhandled error code Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] CDB: cdb[0]=0x28: 28 00 09 20 db b8 00 00 08 00 Aug 4 12:57:23 Tower kernel: end_request: I/O error, dev sda, sector 153148344 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Unhandled error code Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] CDB: cdb[0]=0x28: 28 00 09 21 17 b8 00 00 20 00 Aug 4 12:57:23 Tower kernel: end_request: I/O error, dev sda, sector 153163704 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] READ CAPACITY(16) failed Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Sense not available. Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] READ CAPACITY failed Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Sense not available. Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Write Protect is on Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Mode Sense: 80 80 52 f4 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Got wrong page Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Assuming drive cache: write through Aug 4 12:57:23 Tower kernel: sda: detected capacity change from 1000204886016 to 0 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] READ CAPACITY(16) failed Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Sense not available. Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] READ CAPACITY failed Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Result: hostbyte=0x04 driverbyte=0x00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Sense not available. Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Write Protect is off Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Mode Sense: 00 00 00 00 Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Asking for cache data failed Aug 4 12:57:23 Tower kernel: sd 1:0:0:0: [sda] Assuming drive cache: write through

I doubt that 2 re-allocated sectors would qualify you for an RMA. They'd laugh at you since modern drives of that size have several thousand spare sectors, and your SMART report "normalized" value of 200 is nowhere near the failure threshold of "0" In the same way, the read error rate of "200" is nowhere near its failure threshold. All drives have read errors, some report it, some do not, most will re-try several times before determining a sector is un-readable. As far as IDE vs SATA there is a second setting in your BIOS you need to set to get the drive out of the "legacy" mode. It is probably the root cause of the difference in performance. Joe L.

They all look fine.

Other than you spelled "finished" with a double "n", yes.

As you said, both disks look good. Good luck with your server. Joe L.

Timing, voltage, and clock speed. I would be inclined to believe memtest. And 6-6-6-15 might be too aggressive if the strips are designed for 9-9-9-24 Joe L.

Kernel Oops are usually memory related. Have you run a memory test? Did you specifically set the memory voltage, timing, and clock speed, or did you let the BIOS do it for you. Some BIOS get it right, others get it very wrong. It must be set for your specific model/brand of RAM. How much RAM do you have? Is it less than 512 Meg? Joe L.

I plan on putting a full sized air-filter in front of the fans. I've just not yet built the holder for it. (and even though I have no kids to stop the blades with the toys, fingers, etc... I have managed to get my own fingers in there once or twice ) I plan on using the same style filter as used in my home air-conditioning ducts. the cleaner the air I can force over the disk drives the better. The disks are running very cool. Just sitting there spinning, but not being accessed they are at about 27C, and when being read or written to they get up to 30C. Ambient room air temp is about 25C. When they spin down, their temperatures drop to 25C, same as ambient room air. Joe L.

Time for a few pictures of my second unRAID server. I started with this bargain case: http://www.geeks.com/details.asp?invtid=ET-RM5101-WH&cat=CAS It was $27.99, with free shipping since I ordered some other things at the same time. I took my Dremil tool to the front of it, and completely cut away the center support. I then used aluminum angle to make a pair of drive mounting rails. the holes are spaced so there is about 1/8th of an inch between the drives if I fully populate the case. (it has room on the mounting rail for 14 drives) The rails were spaced so the whole front of the enclosure would become an air plenum for cooling. The drives can very easily be added or removed by opening the front door and unscrewing the two screws. The new drive mounting rails look like this: The power and data connectors are across the top of the server, also very easy to access. Here is a view from the inside, with a few old drives installed to verify spacing. For now I've only have 6 drives connected and configured. The one on the far left is still one of the old scrap drives. I have 3 120mm fans attached to the front door. When it is closed they force air into the cabinet and past the disk drives. I must have gone through about 20 of the small cutting disks with my Dremil tool, but I managed to cut three very nice holes in the front door of the enclosure. It is made of fairly thick steel. (The web-site said 1.2mm steel. That is roughly 18 Gauge) It took a few hours. (And I wore a full face shield, as the thin cutting disks have a tendency to fly apart if you catch them in the steel) I still need to add a holder for an air filter in front of the fans and build a pwm controller adapter to control their speed from the MB connectors. (They are three wire fans, the C2SEE MB I'm using has 4-pin PWM connectors) then, I can make the wiring inside a lot neater by making a custom power harness. The front now looks like this: I still need to get longer SATA cables for the disks, because the case is so deep, the normal 15" ones don't make for a neat cabling job. The same for the power connectors. Even though the power supply has very long cable harnesses, they are not long enough for the SATA connectors closest to the supply to reach the front of the case. I'll probably eventually add a second set of drive support rails for additional drives behind the first. I've easily got the depth in the case. Those first connectors can reach them. That will then let me easily mount another 10 or so drives if I wish. (It will then be my very own budget 24 drive case) Not too bad for a $27.99 case, shipping included. It has been online for about a week now. I've got 3 2TB Hitachi 7200 RPM drives, and 3 1.5TB Seagate 7200 RPM drives installed. The first few days were spend exercising the disks with my preclear_disk script. Then multiple parity calc/parity check runs followed before I added any data. I used "rsync" to copy my entire original unRAID server's data to the new server. The copy with parity enabled, to /mnt/user averaged between 30 and 33 MB/s on larger files. (the copy duplicated my user-shares) right now, it is tower2. Soon, I'll switch their names and my older original server will get a rebuild with the bargain PCI-X based motherboard I picked up recently from geeks.com for $57. (It came with 1Gig of Error-Correcting RAM, Dual Intel Xeon CPUs, heat-sinks, and hardware) I love a bargain. Joe L.

The two files you posted seem to be identical... Both have the same disclaimer at the top of the final output of the pre-clear results: note, some 'raw' values may change, but not be an indication of a problem Only a handful of "raw" values mean anything to anyone but the manufacturer. They intend for us to look only at the "normalized" values. There is one line in the resulting report where a "raw" count changed that we can interpret directly, the line for Re-Allocated Sectors: The pre preclear values were: 5 Reallocated_Sector_Ct 0x0033 100 100 010 Pre-fail Always - 61 The current normalized value is 100 The worst normalized value is 100 The failure threshold for this parameter is 10. ( If the current value drops below the threshold, the manufacturer will consider the drive to have failed.) The number of re-allocated sectors is 61. (This is one of the few parameters where the "raw' value is meaningful to anyone other than the manufacturer) The post preclear values are: 5 Reallocated_Sector_Ct 0x0033 100 100 010 Pre-fail Always - 62 The current normalized value is 100 (it is unchanged from the pre test) The worst normalized value is 100 The failure threshold for this parameter is 10. ( If the current value drops below the threshold, the manufacturer will consider the drive to have failed.) The number of re-allocated sectors is 62. It increased by 1 during the pre-clear process. (This is one of the few parameters where the "raw' value is meaningful to anyone other than the manufacturer) So.. the clearing process found one sector that it could not read in the pre-read phase and was subsequently re-allocated in the clearing of the drive. Typically, in a large drive, there are several thousand spare sectors. The normalized value is unchanged in the post report, and the manufacturer could potentially reject an RMA as the "worst" value is unchanged from its starting value of 100 and nowhere near the failure threshold. If you have time, run another pre-clear cycle or two. If the re-allocated sector count continues to increase, RMA the drive. If it stays stable, just keep an eye on it over the next few months as you use it in your array. Remember that every time you write to the drive the sector will first be read and then written. It makes it easy for the drive to re-allocate an un-readable sector so your data is pretty safe. The pre-clear script has read every sector on the disk at least twice, so you can have some faith in the drive as workable. Joe L.

cache_dirs writes nothing to the disks, your PC might be scanning for new files, If parity spun up, then one of your disks was written to. Perhaps it is making thumbnail images. Joe L.

how much RAM do you have in your server? I know I've run out even with a swap file defined with cache pressure of 10.

I've run out of memory with cache_pressure set to "0" on my server with 512Meg of RAM. It is for that reason I set it to 10. (and it is still possible to run out of RAM if I try hard enough) Default value on Linux is 100. Just use -p 0 to keep your alternative value from being changed

Looks more like you are running out of memory. cache_dirs attempts to set a server kernel variable to be more aggressive in keeping directory entries in cache memory. As part of its script, it invokes sysctl vm.vfs_cache_pressure=10 You can override this with the "-p" option to cache_dirs. As delivered by lime-technology the value of 100 is used. So, you can try adding -p 50 or -p 100 or anything higher than 10. When memory is running out Linux will terminate processes that are idle in order to attempt to free memory for programs requesting it. Typically, the web-management interface is one of those idle, so it gets killed and you lose communications. In the same was SAMBA also gets killed off. You can re-start emhttp with killall emhttp /usr/local/sbin/emhttp & You can re-start samba with /root/samba restart Usually this will allow you to re-gain communications with the server so you can cleanly shut it down and re-start it.

they both look perfectly fine.

You can only exclude top level directories using the "-e" option. You can use the -a option though and add a "prune" command to the "find" being invoked, and it would look like this: cache_dirs -w -B -u -a '-noleaf -name Data -prune -o -print' It was discussed a few post ago. http://lime-technology.com/forum/index.php?topic=4500.msg61520#msg61520 Looking closer at your example, Data is a top level directory. You only need give the share name, not the full path to it. Try this: echo "/boot/scripts/cache_dirs -w -B -u -e Data" | at now + 10 minute also verify the folder does exist with the exact same capitilization as in the directory. type ls -l /mnt/disk*/Data if it does not exist, then perhaps the directory name is lower case and only showing mixed case in Windows when viewed through SAMBA. Joe L.

The argument is passed to the "find" command. Its manual page says You can experiment by changing the number and testing with this sample "find" command and see what it prints: find /mnt/disk*/Music -maxdepth 1 The depth is relative to the starting path, so a maxdepth of 0 would only cache the /mnt/disk*/Music folders and nothing in them or below them. To cache the hierarchy you described you would need at least a depth of 3. Since you have no really deep hierarchies, there is no reason to specify any -d option. There is no penalty of using 5, or using the default of 999. Edit: fixed directory name in example

my fault. You would need to use tee -a /tmp/iso_hb adding the "-a" option to have it append to the same file each time it is invoked, otherwise, it overwrites the file each time. That is why you only see the last file processed. But then, if you run it again, you'll either need to delete the first file contents or you'll see them too with the subsequent runs of handbrake appended.

Thanks Joe L.! Sorry for the late reponse, but I am abroad for a week with very limited access to Internet. I will try out your suggestion when I am back at home. (Concerning the memory issue, I think 3GB should be enough for my test dir with only 3 ISOs , so I guess there has to be something else). Thank you again! with only three, you might agree, BUT are any of the three ISO images over 3GB in size? Remember, they are being read into memory as they are being converted, and the disk buffers will use up nearly all the available memory. As handbrake reads in a .ISO, it might leave far less free memory than you might think. Joe L.